Dynamic Allocation

Dynamic allocation is the core decision-making logic that determines how vault assets are distributed between the Base Layer and Alpha Layer.

Design Philosophy

Every Dawn Vault follows one rule: yield should never drop to zero.

The Base Layer (lending or staking) runs continuously. The Alpha Layer only activates when market conditions make it profitable. This two-layer design means depositors always earn something, with upside capture during favorable periods.

How It Works

Decision Inputs by Vault

Vault	Primary Signal	Secondary Signals
USDC Vault	SOL Funding Rate	Lending rates across protocols
SOL Vault	LST yield − SOL borrow rate spread	Staking APY, borrow rate trends
BTC Vault	SOL FR + USDC borrow cost	BTC price, collateral LTV

USDC Vault Allocation Model

The USDC Vault uses a state machine with two states:

BASE_ONLY ←→ BASE_PLUS_DN

State Transitions

From	To	Condition
BASE_ONLY	BASE_PLUS_DN	SOL FR > 15% annualized for 2 consecutive days
BASE_PLUS_DN	BASE_ONLY	SOL FR < -2% annualized for 1 day
BASE_PLUS_DN	EMERGENCY_EXIT	SOL FR < -10% annualized (no time condition)

Allocation Ranges

State	Lending	Delta-Neutral	Liquidity Buffer
BASE_ONLY	100%	0%	Included in lending
BASE_PLUS_DN	30–50%	50–70%	30% minimum maintained

SOL Vault Allocation Model

Monthly decision cycles to avoid swap cost erosion:

Spread Level	Staking	LST Loop
> 3% (1 week)	30–50%	50–70%
0.5–3%	80–90%	10–20%
< 0% (2 weeks)	100%	0%

BTC Vault Allocation Model

Requires two conditions to be met simultaneously:

1. USDC Vault DN entry conditions satisfied (SOL FR > threshold)

2. USDC borrow APR < SOL FR × 0.6 (borrow cost leaves sufficient margin)

Plus continuous BTC price monitoring for collateral LTV management.

Parameter Optimization

All threshold parameters are:

• Backtested against 5.5 years of historical data

• Stress-tested against major market events

• Externalized as vault configuration (adjustable without code changes)

• Continuously updated based on live performance data

See USDC Vault — Performance for backtest methodology and results.